Working Paper: CEPR ID: DP8626
Authors: Bartosz Adam Mackowiak; Mirko Wiederholt
Abstract: Why were people so unprepared for the global financial crisis, the European debt crisis, and the Fukushima nuclear accident? To address this question, we study a model in which agents make state-contingent plans - think about actions in different contingencies - subject to the constraint that agents can process only a limited amount of information. The model predicts that agents are unprepared in a state when the state has a low probability, the optimal action in that state is uncorrelated with the optimal action in normal times, and actions are strategic complements. We then compare the equilibrium allocation of attention to the efficient allocation of attention. We characterize analytically the conditions under which society would be better off if agents thought more carefully about optimal actions in rare events.
Keywords: disasters; efficiency; rare events; rational inattention
JEL Codes: D83; E58; E60
Edges that are evidenced by causal inference methods are in orange, and the rest are in light blue.
| Cause | Effect |
|---|---|
| agents' limited information processing capabilities (D83) | agents are unprepared for low-probability states (D80) |
| agents are unprepared for low-probability states (D80) | suboptimal actions in low-probability states (D80) |
| suboptimal actions in low-probability states (D80) | expected loss from suboptimal actions is disproportionately higher (D80) |
| independence of optimal actions across states (C61) | quality of decisions (D79) |
| if optimal actions are uncorrelated (C69) | worse actions in low-probability states (D80) |
| strategic complementarity in actions (C72) | allocation of attention (D91) |
| agents less likely to prepare for unusual times if others are not preparing (D84) | neglect planning for rare events (D81) |